On Tue, Dec 12, 2017 at 02:24:54PM +0100, Léo FERLIN SUTTON wrote: > We believe this is related to the size of our pg_catalog, we have > thousands (if not hundred of thousands of views) on most of our > databases. Here is an example on cluster A : ... > We are completely aware that this is a *bad* idea, and our newer I appreciate your candor. :-) > developments are all moving far far away for this type of schema, > however we are currently stuck with this for at least a few more > months if not years. > > My question to this mailing list is : Are we missing something that > could speed up the pg_dump ? There isn't much we can do to speed it up anymore. We optimized the schema restore in the past, at least with the ideas we had. A new idea we discussed last month was to allow the schema dump/restore while the old server is running. You can see the discussion here: https://www.postgresql.org/message-id/flat/20171205140135.GA25023%40momjian.us#20171205140135.GA25023@xxxxxxxxxx That is only brainstorming at this point and it is not clear we will even implement it. The bottom line is that pg_upgrade is much more effective on large databases with a smaller number of database objects than the reverse. -- Bruce Momjian <bruce@xxxxxxxxxx> http://momjian.us EnterpriseDB http://enterprisedb.com + As you are, so once was I. As I am, so you will be. + + Ancient Roman grave inscription +
